iec 61508 e .biec/en 61508. the safety life cycle of the hardware, the architecture requirements the...

© by HIMA Paul Hildebrandt GmbH + Co KG 1

Safety considerations

Dr. Josef Börcsök, HIMA Paul Hildebrandt GmbH + Co KG

A considerable amount of data is required in order to be able to assess safety systems properly. One of the most important criteria is consideration of the distribution of failures over a system’s life cycle. In considering such failures, a basic distinction is made between safe and dangerous failures. Safe failures are further divided into

• safe detectable, and • safe undetectable

failures. Safe failures, whether detected or undetected, are failures that exert no influence on the safe operation of the system. This is not the case with dangerous failures. These failures, when they occur, lead to a hazardous situation in the system which may even, under certain circumstances, seriously endanger human life. These failures are also divided into

• dangerous detectable, and • dangerous undetectable

failures. In the event of dangerous detectable failures, however, the safety system, provided it is appropriately designed, can bring the entire system or plant into a safe state. It is undetectable, dangerous failures that constitute a critical state. No safety system is able to detect such failures when they occur. They may be present in the system until it switches off or, in the worst-case scenario, until it fails dangerously without the user being aware of it. HIMA systems are always developed, produced and certified in accordance with the prevailing national and international standards. One of the most important international standards in this regard is IEC/EN 61508. IEC/EN 61508 covers not only pure arithmetical values, such as PFD and PFH, which provide information about the probability of system failure, but also a system’s entire safety life cycle.


Figure 1: Representation of the safety life cycle Consideration of the safety life cycle facilitates a systematic approach to the problems of functional safety. Moreover, the SIL capability each individual safety function must have is also laid down here (Table 1). Table 1: SILs for low and high demand modes of operation Safety integrity level (SIL)

Low demand mode of operation (mean probability of failure to perform design function on demand)

High demand mode of operation (probability of a dangerous failure per hour)

4 ≥10-5 to


Specification of the demands on hardware is an important part of IEC/EN 61508. The safety life cycle of the hardware, the architecture requirements as well as type A (whose behaviour in the event of failure is fully known) and type B (whose behaviour in the event of failure is not fully known) subsystems and the corresponding SFF (safe failure fraction) are also defined here. In order to draw up a specification of the safety function, precise information is required on how the required safety is to be achieved and maintained. Table 2: Type A subsystems and Type B subsystems

Type A Type B Hardware failure tolerance Hardware failure tolerance Safe failure

fraction 0 failures 1 failure 2 failures 0 failures 1 failure 2 failures< 60 % SIL 1 SIL 2 SIL 3 Not

allowed SIL 1 SIL 2

60 % - < 90% SIL 2 SIL 3 SIL 4 SIL 1 SIL 2 SIL 3 90 % - < 99% SIL 3 SIL 4 SIL 4 SIL 2 SIL 3 SIL 4

≥ 99 % SIL 3 SIL 4 SIL 4 SIL 3 SIL 4 SIL 4 A safety-related system must be designed in accordance with the safety specification drawn up for it. The hardware architecture requirement has to be matched to the required SIL capability. This safety integrity level is limited by the hardware failure tolerance and the share of non-dangerous errors (Table 2). For example, the equations for the PFD/PFH calculations of different HIMA systems should be given here. In contrast to many other reports currently circulating, these observations are based on the applicable equations in IEC/EN 61508 and relate to a 10-year time period. Many reports are to be found that relate to a time period of half a year and are based on the simplified ISA equations but passed off as the IEC/EN 61508 figures. These calculations do not include common-cause failures or the system’s diagnostic coverage capability. This gives rise to sometimes considerable discrepancies in the PFD figures. These consequences are always taken into account in all calculations and certification processes relating to HIMA systems, which consequently conform fully to IEC/EN61508. This report uses various examples of actual values to illustrate the calculation of the safety integrity level of various HIMA systems. First, however, the individual IEC/EN 61508 equations for the various system architectures must be presented in order to make a clear distinction between them and the ISA norm.


Equation for determining PFD of a 1oo1 system:

MTTR)MTTR2T

tt)(PFD

DD1

DU

CED

CEDDDU1oo1,G

⋅+

+⋅=

⋅=

⋅+=

λλ

λλλ

(1)

with

MTTR)MTTR2T

tD

DD1

D

DUCE ⋅+

+⋅=

λλ

λλ (2)

Equation for determining PFH of a 1oo1 system:

DU1oo1,GPFH λ= (3) Equation for determining PFD of a 1oo2 system:

( ) ( )( )

+⋅⋅+⋅⋅+

⋅⋅⋅−+⋅−⋅=

MTTR2T

MTTR

tt112PFD

1DUDDD

GECE2

DUDDD2oo1,G

λβλβ

λβλβ (4)

with

MTTR)MTTR2T

tD

DD1

D

DUCE ⋅+

+⋅=

λλ

λλ (5)

MTTR)MTTR3T

tD

DD1

D

DUGE ⋅+

+⋅=

λλ

λλ (6)

Equation for determining the PFH of a 1oo2 system:

( ) ( )( )DUDDD

CE2

DUDDD2oo1,G t112PFHλ⋅β+λ⋅β+

⋅λ⋅β−+λ⋅β−⋅= (7)

with

MTTR)MTTR2T

tD

DD1

D

DUCE ⋅+

+⋅=

λλ

λλ (8)


Equation for determining the PFD of a 2oo3 system:

( ) ( )( )

+⋅⋅+⋅⋅+

⋅⋅⋅−+⋅−⋅=

MTTR2T

MTTR

tt116PFD

1DUDDD

GECE2

DUDDD3oo2,G

λβλβ

λβλβ (9)

with

MTTR)MTTR2T

tD

DD1

D

DUCE ⋅+

+⋅=

λλ

λλ (10)

MTTR)MTTR3T

tD

DD1

D

DUGE ⋅+

+⋅=

λλ

λλ (11)

Equation for determining the PFH of a 2oo3 system:

( ) ( )( )DUDDD

CEDUDDDooG tPFHλβλβ

λβλβ⋅+⋅+

⋅⋅−+⋅−⋅= 232, 116 (12)

with

MTTR)MTTR2T

tD

DD1

D

DUCE ⋅+

+⋅=

λλ

λλ (13)

Two further important indicators for safety-related systems are the SFF (safe failure fraction) and the DC (diagnostic coverage) factor. The SFF can be calculated by means of the following equation:

DUDDS

DDSSFFλλλ

λλ++

+= (14)

The DC factor can be determined by means of the following equation:

∑∑=

D

DDDCλ

λ (15)


The SFF represents the share of safety-relevant failures and the DC factor the level of diagnostic coverage. The meaning of the individual factors in these equations is as follows: β weighting factor for dangerous, undetected, common-cause failures Dβ weighting factor for dangerous, detected, common-cause failures Dλ system failure rate due to dangerous failures DDλ system failure rate due to dangerous, detected failures DUλ system failure rate due to dangerous, undetected failures

MTTR mean time to repair PFDG average probability of failure on low demand PFHG average probability of failure on high demand T1 proof test time tCE average channel-related failure time tGE average system-related failure time In order to determine the safety integrity of safety-related systems consisting of several individual systems, the mean probability of failure PFDsys or PFHsys is required for the whole system. PFDsys or PFHsys is determined by obtaining and summing the mean probabilities for the individual systems.

FELSsys PFDPFDPFDPFD ++= (16) or

FELSsys PFHPFHPFHPFH ++= (17) In order to determine the mean probability for each part of the system, the following information must be available:

• the basic architecture • the diagnosis coverage of each channel • the failure rate per hour for each channel • the factors β and βD for common-cause failures


This last list introduces the notion of ‘common-cause failures’. The aim here is to detect common-cause failures as early as possible and to bring the system into a safe state. The β-factor is introduced as the fraction of common-cause failures to normal (single) failures. In the following, various system architectures are presented and the PFD/PFH values for the individual systems established. The following parameters should be identical for all systems β D = common-cause factor for detectable failures β = common-cause factor for undetectable failures T1 = maintenance interval MTTR = mean time to repair And have the following values β D = 1% β = 2% T1 = 10 years MTTR = 8 hours


The following individual systems are used in various configurations in the example calculations:

For all the following architectures, the following two points should always apply:

• sensors in 2oo3 architecture • actuators in 1oo2 architecture

Module Pressure sensorTemp.sensor

DI:F 3236

DI: F 3238

AI: F 6214

AI:F6217

V-BG: F 7553

CPU: F 8650E

DO:F 3330

DO: F 3334

AO: F 6705

Actuator: valve

lambda_b in [1/h] 7,43E-07 1,09E-06 1,11E-06 1,11E-06 5,60E-07 2,08E-06 8,17E-07 6,21E-07 9,45E-07

MTTF in [years] 153,66 104,60 102,80 103,17 203,99 54,79 139,78 183,91 120,79

Proof-checkinterval T1in [years]

10 10 10 10 10 10 10 10 10 10 10 10

MTTR in [h] 8 8 8 8 8 8 8 8 8 8 8 8

βD 0,01 0,01 0,01 0,01 0,01 0,01 0,01 0,01 0,01 0,01 0,01 0,01

β 0,02 0,02 0,02 0,02 0,02 0,02 0,02 0,02 0,02 0,02 0,02 0,02

PFD1oo1in [1]

9,79E-06 2,38E-05 5,12E-05 2,87E-05 9,78E-06 2,94E-05 8,18E-06 1,40E-05 1,86E-05

PFH1oo1 in [1/h]

2,05E-10 5,14E-10 1,11E-09 9,64E-10 5,76E-10 4,08E-09 6,58E-10 6,87E-10 6,17E-10

PFD1oo2/2oo3 in [1] *)

1,00E-04 1,56E-04 3,33E-05

PFH1oo2/2oo3in [1/h] *)

2,22E-08 3,47E-08 7,40E-09

TÜV claimed SIL 3 3 3 3 3 3 3 3 3


The following systems are to be investigated: System 1 (Digital Loop 1)

Pressure sensor(2oo3)

Temperaturesensor (2oo3)

DIF 3238

DIF 3238

V-BGF 7553

CPUF 8650E

DOF 3334

Actuator: Valve (1oo2)

Architecture PFD-IEC in [1]PFH-IEC in [1/h]

SIL, PFD-IEC

SIL, PFH-IEC

Pressure sensor 2oo3 1.00E-04 2.22E-08 3 3Temp. sensor 2oo3 1.56E-04 3.47E-08 3 3DI: F 3238 1oo2 4.63E-07 1.55E-09 4 3V-BG: F 7553 1oo1 9.78E-06 5.76E-10 4 4CPU: F 8650E 1oo1 2.94E-05 4.08E-09 3 3DO: F 3334 1oo1 1.40E-05 6.87E-10 4 4Actuator: valve 1oo2 3.33E-05 7.40E-09 4 3

System without sensor or actuator 5.36E-05 6.90E-09 4 4

System with sensor and actuator 3.43E-04 7.12E-08 3 3

TÜV claimed SILfor system without sensor and actuator

3 3


System 2 (Digital Loop 2)



DIF 3238

DIF 3238

V-BGF 7553

CPUF 8650E

DOF 3334

Actuator: valve (1oo2)

CPUF 8650E


SIL, PFD-IEC

SIL, PFH-IEC




TÜV claimed SILfor system without sensor or actuator

3 3





DIF 3238

DIF 3238

V-BGF 7553

CPUF 8650E

DOF 3334


DIF 3238

DOF 3334


SIL, PFD-IEC

SIL, PFH-IEC





3 3





DIF 3238

DIF 3238

V-BGF 7553

CPUF 8650E

DOF 3334


CPUF 8650E

DIF 3238

DOF 3334


SIL, PFD-IEC

SIL, PFH-IEC





3 3


System 5 (Analog-Digital Loop 1)



AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

DOF 3334



SIL, PFD-IEC

SIL, PFH-IEC

Pressure sensor 2oo3 1.00E-04 2.22E-08 3 3Temp. sensor 2oo3 1.56E-04 3.47E-08 3 3AI: F 6214 1oo2 1.00E-06 3.44E-09 4 3V-BG: F 7553 1oo1 9.78E-06 5.76E-10 4 4CPU: F 8650E 1oo1 2.94E-05 4.08E-09 3 3DO: F 3334 1oo1 1.40E-05 6.87E-10 4 4Actuator: valve 1oo2 3.33E-05 7.40E-09 4 3




3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

DOF 3334


CPUF 8650E


SIL, PFD-IEC

SIL, PFH-IEC





3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

DOF 3334


AIF 6214

DOF 3334


SIL, PFD-IEC

SIL, PFH-IEC





3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

DOF 3334


CPUF 8650E

AIF 6214

DOF 3334


SIL, PFD-IEC

SIL, PFH-IEC





3 3


System 9 (Analog Loop 1)



AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

AOF 6705



SIL, PFD-IEC

SIL, PFH-IEC

Pressure sensor 2oo3 1.00E-04 2.22E-08 3 3Temp. sensor 2oo3 1.56E-04 3.47E-08 3 3AI: F 6214 1oo2 1.00E-06 3.44E-09 4 3V-BG: F 7553 1oo1 9.78E-06 5.76E-10 4 4CPU: F 8650E 1oo1 2.94E-05 4.08E-09 3 3AO: F 6705 1oo1 1.86E-05 6.17E-10 3 4Actuator: valve 1oo2 3.33E-05 7.40E-09 4 3




3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

AOF 6705


CPUF 8650E


SIL, PFD-IEC

SIL, PFH-IEC




TÜV claimed SILfor system without system or actuator

3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

AOF 6705


AIF 6214

AOF 6705


SIL, PFD-IEC

SIL, PFH-IEC





3 3





AIF 6214

AIF 6214

V-BGF 7553

CPUF 8650E

AOF 6705


CPUF 8650E

AIF 6214

AOF 6705


SIL, PFD-IEC

SIL, PFH-IEC





3 3


Bibliography

[1] IEC/EN 61508: International Standard 61508 Functional Safety: Safety-Related System. Geneva, International Electrotechnical Commission [2] Börcsök, J.: IEC/EN 61508- eine Norm für viele Fälle, atp 44, 2002 Oldenbourg-Verlag [3] Börcsök, J.: Konzepte zur methodischen Untersuchung von Hardwarearchitekturen in sicherheitsgerichteten Anwendungen, erscheint im VDE–Verlag 2003 [4] Börcsök, J.: Sicherheits-Rechnerarchitekturen Teil 1 und 2, Vorlesung Universität Kassel 2000/2001 [6] DIN VDE 0801: Funktionale Sicherheit sicherheitsbezogener elektrischer/elektronischer/programmierbarer elektronischer Systeme (E/E/PES), (IEC 65A/255/CDV:1998), S. 27f, August 1998. [7] DIN V 19250: Grundlegende Sicherheitsbetrachtungen für MSR- Schutzeinrichtungen. Beuth Verlag Berlin 1998 [8] DIN VDE 0801/A1: Grundsätze für Rechner in Systemen mit Sicherheitsaufgaben. Beuth Verlag [9] IEC 60880-2: Software für Rechner mit sicherheitskritischer Bedeutung. 12/2001

iec 61508 e .biec/en 61508. the safety life cycle of the hardware, the architecture requirements the...

Documents